ResearchTasksMultimodal Reasoning AverageFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedCombined (M3-Bench + Video-MME-Long)M3-Agent + StreamMeCo + TMR46.8Average Score24Apr 13, 2026